EP1229576B1

EP1229576B1 - Method of producing SOI MOSFET

Info

Publication number: EP1229576B1
Application number: EP02002458A
Authority: EP
Inventors: Albert Oscar Adan
Original assignee: Sharp Corp
Current assignee: Sharp Corp
Priority date: 2001-02-02
Filing date: 2002-02-01
Publication date: 2008-06-11
Anticipated expiration: 2022-02-01
Also published as: DE60227019D1; KR100466061B1; TW544844B; US6627505B2; JP3531671B2; CN1369903A; KR20020064674A; JP2002231960A; EP1229576A2; CN1194395C; US20020177286A1; EP1229576A3

Description

BACKGROUND OF THE INVENTION

1.Field of the Invention

The present invention relates to a method of producing a SOI (silicon-on-insulator) MOSFET (metal oxide semiconductor field effect transistor), more particularly, to a method of producing a SOI MOSFET with reduced electrical characteristic fluctuations related to variations in the thickness of a top semiconductor layer (i.e., an active semiconductor layer) and a MOS FET.

2.Description of Related Art

Generally known MOSFETs formed on SOI substrates such as SOS (silicon on sapphire), SIMOX (silicon separation by ion implantation of oxygen) and BSOI (bonded SOI) substrates have advantages in low-voltage and high-speed operation. In addition to that, the SOI MOSFETs have an advantage over devices formed on bulk silicon substrates in that layout area for the SOI MOSFETs is smaller.
However, the SOI MOSFETs have only three terminals (gate, drain and source) while bulk silicon MOSFETs have four terminals (gate, drain, source and substrate). For this reason, the SOI MOSFETs are inferior in electrical characteristics, especially, short channel effect, drain/source blocking voltage, punch through and the like.
Referring to Figs. 10 (a) and 10(b), in a bulk silicon MOSFET, the base terminal of a parasite bipolar (NPN) transistor is tied to the substrate and the substrate-source junction is reversely biased. As a result, if an impact ion current Ii is generated near a drain region, the parasite bipolar transistor has very little effect on operation of the MOSFET.
In contrast, referring to Figs. 9(a) and 9(b), in a SOI MOSFET, the base terminal of a parasite bipolar transistor is a top semiconductor layer in a floating state. As a result, in usual operation, an impact ion current Ii generated near a drain region acts as a base current of the parasite bipolar transistor to generate a positive feed-back effect, which results in reduction in the short-channel effect and decrease in the drain/source blocking voltage. In the case where a channel region is formed in a relatively thick top semiconductor layer, the channel region behaves in a partially depleted mode and a so-called kink effect is produced in output characteristics owing to impact-ionization. Therefore, the electrical characteristics of the SOI MOSFET are significantly affected.
Here, the kink is a phenomenon in which majority carriers generated by impact ionization accumulate to raise the potential of the floating substrate, bring down threshold voltage and further cause drain current to increase abruptly. Thus the operation of the SOI MOSFET is greatly affected.
In order to realize a fully depleted SOI free of the kink effect, there is a technique of forming a top silicon layer which is thinner than a depletion layer induced by the gate electrode. Generally, as shown in Fig. 11, the full depletion of the top silicon layer requires adjustment of the thickness of the top silicon layer and the impurity concentration Na in the substrate.
However, as understood from Fig. 11, a major drawback of the fully depleted SOI transistor is that the threshold voltage Vth is sensitive to the thickness of the top silicon layer.
That is, the threshold is represented by $\begin{array}{l} Vth & ≅ V_{fbt} + \frac{q \cdot Na \cdot Tsi}{C_{tox}} + 2 \cdot Φ_{F} - \frac{C_{box}}{C_{tox}} (V_{sub} - V_{fbb}) \\ \frac{Δ Vth}{Δ Tsi} & \approx \frac{q \cdot Na}{C_{tox}} . \end{array}$

wherein V_fbt is flat band voltage (at the top of the top silicon layer), V_fbb is flat band voltage (at the bottom of the top silicon layer), C_tox is capacity of a gate insulating film, Na is impurity concentration in the substrate, T_si is thickness of the top silicon layer, φ F is Fermi potential and V_sub is substrate voltage. For typical values of Na and the thickness of the gate insulating film, Δ Vth/ Δ T_si is about 10 mV/nm.
The threshold voltage affects or is related to electrical parameters such as OFF-state current which exponentially depends on the threshold voltage, as shown in the following formula: $I_{doff} ≅ Io \cdot W \cdot 10^{(- \frac{Vth}{S})}$

wherein W is channel width of the transistor and I₀ is a constant when the gate voltage is 0V (I₀= about 10^-7A/µm).
For example, with a fully depleted SOI transistor (slope S (S factor) in a sub-threshold region is about 65 mV/dec), the OFF current varies 10 times if the threshold voltage changes by 65 mV. Thus, to control the threshold voltage is important for the characteristics of the semiconductor device.
In 1995 IEEE International SOI Conference Short Course, there was proposed a method for suppressing variations in the threshold of the SOI MOSFET using a constant dose method. In the constant dose method, ion implantation is carried out on the top silicon layer of the SOI substrate under such conditions that the dose D = Na × T_si is constant. As a result, it is understood from Formula (1) that the change of the threshold voltage Vth with respect to the thickness T_si of the top silicon layer is suppressed. This is also clear from the relationship of T_si to Vth shown in Fig. 12.
To suppress the dependency of Vth on Tsi, there is proposed a method of combining a partially depleted SOI and a fully depleted SOI as shown in Fig. 13 (Japanese Unexamined Patent Publication HEI 6(1994)-268215 ). In this device, the impurity concentration is higher at a channel edge 11 than at channel center 12 in the top silicon layer, and thereby, the channel edge 11 is not fully depleted but the channel center 12 is fully depleted. As a result, the threshold voltage of the SOI transistor is determined by the impurity concentration at the channel edge, and this device operates as a partially depleted device.
USP 5841170 discloses a SOI MOSFET whose channel region has an impurity profile that is nonuniform in a source/drain direction. This device is designed to have impurity concentrations such that full depletion is realized both at the channel center and at the channel edge. Thus, this device operates in a fully depleted mode and prevents the kink effect.
However, although the dependency of Vth on the thickness of the top silicon layer can be reduced by the constant dose method, the short-channel effect and the punch through are not considered. Since the impurity implantation is uniformly performed over the channel, the short-channel effect is more easily induced to take place.
The structure of the device proposed by Japanese Unexamined Patent Publication No. HEI 6(1994)-268215 can reduce the short-channel effect, but since the device operates in a partially depleted mode, the device is more susceptible to the kink effect and floating substrate effect.
The device proposed by USP 5841170 does not give any consideration to fluctuations in electrical characteristics related to variations in the thickness of the top silicon layer.
US 5.656.844 discloses a semiconductor-on-insulator transistor having a channel region in a semiconductor film under a gate insulating-layer. The channel region has a top dopant concentration N_T at a top surface of the film that is significantly greater than a bottom dopant concentration N_B at a bottom surface of the film. This non-uniform doping profile provides an SOI device that operates in a fully-depleted mode yet permits thicker films without a significant degradation of sub-threshold slope.

SUMMARY OF THE INVENTION

The invention is defined in independent claims 1 and 2.
A preferred embodiment provides a method of producing a SOI MOSFET which includes a fully depleted channel region of a first conductivity type formed in a top semiconductor layer disposed on an insulative substrate, source/drain regions of a second conductivity type formed to sandwich the channel region and a gate electrode formed on the channel region with intervention of a gate insulating film, the method comprising:

forming the channel region by setting an impurity concentration of channel edge regions of the channel region adjacent to the source/drain regions higher than an impurity concentration of a channel central region of the channel region, and setting a threshold voltage Vth₀ of the channel central region and a threshold voltage Vth_edge of the channel edge regions so that a change of the threshold voltage Vth₀ with respect to a change of the thickness of the top semiconductor layer and a change of the threshold voltage Vth_edge with respect to the change of the thickness of the top semiconductor layer are of opposite sign.

That is, in view of the above-described problems, an object of the present invention is to provide a method of producing a highly reliable SOI MOSFET by effectively reducing the short-channel effect, the punch through and the like and suppressing the influence of the thickness of the top semiconductor layer on the electrical characteristics of the SOI MOSFET, and such a highly reliable SOI MOSFET.
These and other objects of the present application will become more readily apparent from the detailed description given hereinafter. However, it should be understood that the detailed description and specific examples, while indicating preferred embodiments of the invention, are given by way of illustration only, since various changes and modifications within the scope of the appended claims will become apparent to those skilled in the art from this detailed description.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a schematic sectional view of a major part of a SOI MOSFET illustrating an embodiment in accordance with the present invention;
Figs. 2(a) and 2(b) are schematic sectional views of a major part of a SOI MOSFET illustrating impurity concentration profile in the channel region of the SOI MOSFET in accordance with the present invention;
Fig. 3 is a graphical representation showing a relationship between the thickness T_si of the top silicon layer and threshold voltage Vth;
Figs. 4(a) to 4(c) are schematic sectional views of a SOI MOSFET illustrating production steps for producing the SOI MOSFET in accordance with the present invention;
Fig. 5 is a graphical representation showing a relationship between the projected range of impurity ions/the thickness of the top silicon layer and the threshold voltage;
Fig. 6 is a graphical representation showing relationships between the thickness T_si of the top silicon layer and threshold voltage Vth at varied ion implantation energies;
Fig. 7 is a graphical representation showing a relationship between the projected range of impurity ions/the thickness of the top silicon layer and the change of the threshold voltage/the change of the thickness of the top silicon layer;
Figs. 8(a) and 8(b) are graphical representations showing relationships between the thickness of the top silicon layer and the threshold voltage Vth;
Figs. 9(a) and 9(b) are a schematic sectional view of a prior-art SOI MOSFET and its equivalent circuit diagram;
Figs. 10(a) and 10(b) are a schematic sectional view of a prior-art MOSFET and its equivalent circuit diagram;
Fig. 11 is a graphical representation showing relationships between the thickness T_si of the top silicon layer and the threshold voltage Vth;
Fig. 12 is a graphical representation showing relationships between the thickness T_si of the top silicon layer and the threshold voltage Vth in a MOSFET formed by a prior-art constant dose method; and
Fig. 13 is a schematic sectional view of another prior-art MOSFET.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The SOI MOSFET of the present invention is formed on a substrate of SOI structure which is composed of an insulative substrate and a top semiconductor layer, and is mainly composed of a channel region of a first conductivity type, source/drain regions of a second conductivity type and a gate electrode formed on the channel region with intervention of a gate insulating film.
The insulative substrate of the SOI-structure substrate of the invention may be a substrate formed of an intrinsically insulative material such as sapphire, quartz, glass, plastic or the like, or a substrate in which a buried insulating film is formed on a support substrate. Here, examples of the support substrate include a substrate made of an element semiconductor such as silicon, germanium or the like and a substrate made of a compound semiconductor such as GaAs, InGaAs or the like. Among these substrates, a single crystal silicon substrate or a polysilicon substrate is preferred. Examples of the buried insulating film include a single-layer film or a multi-layer film of SiO₂, SiN and the like. The thickness of the insulative substrate may be set as appropriate depending upon the desired characteristics of the semiconductor device to be produced, the voltages to be applied when the produced semiconductor device is used, and others, but may be about 50 to 1,000 nm, preferably about 80 to 500 nm, for example.
The top semiconductor layer is typically a semiconductor thin film which functions as an active layer for forming a transistor and may be formed of a thin film of an element semiconductor such as silicon, germanium or the like or a compound semiconductor such as GaAs, InGaAs or the like. Among these, a silicon thin film is preferred. More preferably, the silicon thin film is formed of single crystal silicon. The thickness of the top semiconductor layer may be set as appropriate depending upon the structure of the semiconductor device to be produced, but may be about 10 to 1,000 nm, preferably about 10 to 500 nm, more preferably about 20 to 70 nm, for example.
The SOI-structure substrate may typically be composed of a support substrate having a buried insulating film and a top semiconductor layer which are formed on the support substrate. However, the SOI-structure substrate may be a multi-layer SOI substrate composed of a support substrate having a first buried insulating layer, a first top semiconductor layer, a second buried insulating layer, a second top semiconductor layer, ... which are formed sequentially on the support substrate. Usable SOI-structure substrates include, for example, a SIMOX (separation by implantation of oxygen) type substrate wherein a semiconductor substrate is implanted with oxygen ions and thermally treated to form a buried oxide film as a first insulating layer in the semiconductor substrate; a BSOI (bonded SOI) substrate wherein two semiconductor substrates having oxide films formed thereon by thermal oxidation are bonded; a SOI substrate wherein a first insulating layer and a first semiconductor layer are formed on a semiconductor substrate by epitaxial growth; a so-called bonded multi-layer SOI substrate formed by bonding a SOI substrate wherein a first insulating layer and a first semiconductor layer are formed on a semiconductor substrate by epitaxial growth to a semiconductor substrate wherein an oxide film is formed on its surface by thermal oxidation or by epitaxial growth; and a multi-layer SOI substrate wherein a first insulating layer, a first semiconductor layer, a second insulating layer and a second semiconductor layer are formed on a semiconductor substrate by epitaxial growth. The SOI-structure substrate may be a substrate on which elements such as transistors, capacitors and the like or circuits are formed and device isolation regions may be optionally formed on the substrate by a LOCOS (local oxidation of silicon) separation method, a trench isolation method, an STI (shallow trench isolation) method or the like. One or more P-type or N-type wells may be formed on the SOI-structure substrate.
The MOSFET in the present invention may be of N-channel type or P-channel type or may be of both types.
The channel region of the first conductivity type of the MOSFET may be of P-type or N-type. The channel region has a channel central region and channel edge region. The channel edge region is located on edge of the channel regions, adjacently to the source/drain regions described later, and have a impurity concentration higher than the impurity concentration of the channel central region. In other words, the channel regions have a non-uniform impurity concentration profile in a source-to-drain direction. Difference between the impurity concentration Nb in the channel central region and the impurity concentration Na in the channel edge regions is not particularly limited so long as the impurity concentrations Na and Nb have a relationship described later, but may be Na/Nb = 3 to 6 approximately.
In the channel central region, the threshold voltage Vth₀ of the channel central region is set at a desired value so that a change of the threshold voltage Vtho is a positive or a negative value with respect to a change of the thickness T_si of the top semiconductor layer. In the channel edge regions, the threshold voltage Vth_edge is set at a desired value so that a change of the threshold voltage Vth_edge of the channel edge regions is a negative or a positive value with respect to the change of the thickness T_si of the top semiconductor layer, that is, the change of the threshold voltage Vth_edge of the channel edge regions is of sign opposite to the change of the threshold voltage Vtho of the channel central region. More particularly the change of the threshold voltage Vtho of the channel central region and the change of the threshold voltage Vth_edge of the channel edge regions preferably meet sign(Δ Vth₀/ Δ T_si) < 0 and sign(Δ Vth_edge/ Δ T_si) > 0, respectively, or sign(Δ Vth₀ / Δ T_si) > 0 and sign(Δ Vth_edge/ Δ T_si) < 0, respectively.
In the SOI MOSFET having such a channel region, the threshold voltage Vth of the channel region as a whole is represented by the following formula: $Vth = {Vth}_{0} + {Vth}_{edge}$

(wherein Vtho is the threshold of the channel central region and Vth_edge is the threshold of the channel edge region). Accordingly, it is preferable that the above-mentioned relations are satisfied and the absolute value of sign(Δ Vth₀ / Δ T_si) is almost equal to the absolute value of sign(Δ Vth_edge/ Δ T_si). In other words, the change of the threshold voltage Vth of the entire channel region with respect to the change of the thickness T_si of the top semiconductor layer preferably meets (Δ Vth/ Δ T_si) ≒ 0. That (ΔVth/ΔT_si) is almost 0 means that the change of the threshold voltage Vth of the entire channel region is almost cancelled out with respect to the change of the thickness of the top semiconductor layer. The change of the threshold voltage Vth of the channel region as a whole is determined by the specification of the device and its production process. Generally, the change of the threshold voltage Vth is represented by the following formula with regard to mutually non-related parameters: $Δ Vth ≅ \sqrt{{({{}^{Δ Vth}/}_{Δ Tsi})}^{2} Δ {Tsi}^{2} + {({{}^{Δ Vth}/}_{Δ Tox})}^{2} Δ {Tox}^{2} + {({{}^{Δ Vth}/}_{Δ L})}^{2} Δ L^{2}}$

In this case, major factors of variations of the change are T_si, L and T_ox. For example, regarding a typical transistor with L = 0.2µm, T_ox = 5 nm and T_si = 50 nm, there are obtained Δ L = ± 0.7µm, ΔT_ox = ± 0.5µm, Δ Tsi/Tsi < 10 % and consequently (Δ Vth/ Δ Tsi) < 1 mV/nm, which is extremely small.
The channel central region has a lateral length of suitably about half, preferably about two-fifths, more preferably one-third of the minimum gate length of the SOI MOSFET. Specifically, about 0.01 µm to 0.4 µm and preferably about 0.03µm to 0.13µm may be mentioned. Preferably, the channel central region has an almost uniform impurity concentration in a depth direction and in a horizontal direction.
The channel edge regions have a lateral length of suitably about half, preferably about two-fifths, more preferably one-third of the minimum gate length of the SOI MOSFET. Specifically, about 0.01µm to 0.4µm and preferably about 0.03µm to 0.13µm may be mentioned. Preferably, the channel edge regions have an almost uniform impurity concentration in the depth direction and in the horizontal direction. That is because, if the channel edge regions have uniform impurity concentration, the threshold voltage Vth_edge changes linearly with respect to the thickness T_si of the top silicon layer. The channel edge regions may have different impurity concentrations and different impurity concentration distributions, but preferably, have the same impurity concentration and impurity concentration distribution.
The source/drain regions of the second conductivity type of the MOSFET are of a conductivity type opposite to the conductivity type of the channel region, and suitably have an impurity concentration of about 1 to 10×10²⁰ ions/cm³. The source/drain regions may be a LDD structure, a DDD structure or the like.
The gate insulating film of the MOSFET, as gate insulating films of usual MOS transistors, may be formed of a single-layer film or a multi-layer film of a silicon oxide film, a silicon nitride film, a highly dielectric film (for example, Ti₂O₅) and the like. The thickness thereof may be about 2 to 7 nm in terms of SiO₂.
The gate electrode of the MOSFET, as gate electrodes of usual MOS transistors, may be formed of polysilicon; a silicide of a high-melting metal such as W, Ta, Ti, Mo or the like; a polycide formed of the above-mentioned silicide and polysilicon; other metal or the like, in a thickness of about 150 nm to 300 nm. The size and the shape of the gate electrode are not particularly limited, and may be selected as appropriate for obtaining desired characteristics. Sidewall spacers may be formed on the sidewalls of the gate electrode.
In order to form the above-described SOI MOSFET, that is, in order to realize an optimal impurity concentration profile in the channel region, impurity ions may be implanted at an implantation energy such that the peak of concentration is positioned at Rp (average projected range) from the surface and Rp/T_si is 0.5 or less, more preferably about 0.25.
The SOI semiconductor device of the present invention may be formed using a general MOS process or CMOS process technique or using such a technique as a base process which may be modified appropriately for producing the above-described SOI semiconductor device. In the method of producing a SOIMOSFET of the present invention, in addition to a step for forming the above-described impurity concentration profile, may be performed as appropriate in optional order the formation of wells in the semiconductor substrate or the semiconductor layer, the formation of well contacts, the impurity introduction into the top semiconductor layer, the formation of the source/drain regions, the formation of LDD regions and/or the like, the formation of sidewall spacers, the formation of interlayer insulating films, the formation of contact holes in the interlayer insulating films, the formation of wiring layers, thermal treatment and the like.
The SOI MOSFET device of the present invention is now explained in detail below.
As shown in Fig. 1, Fig. 2(a) and Fig. 2(b), in the SOI MOSFET of the present invention, N-type source/drain regions 10 reaching a buried insulating film 2 are formed in a top silicon layer 3 of a SOI substrate 4 which is composed of a silicon substrate 1, the buried insulating film 2 and the top silicon layer 3. A gate electrode 8 is formed on the top silicon layer 3 between the source/drain regions 10 with intervention of a gate insulating film 7. Under the gate electrode 8, there is formed a channel region including a central region P2, and edge regions P1 and P3. The region P2 is adjusted to have an impurity concentration Nb (e.g., about 5 × 10¹⁶ ions/cm³), and the regions P1 and P3 are adjusted to have an impurity concentration Na (e.g., about 3 × 10¹⁷ ions/cm³) (see Fig. 2(b)).
In this SOI MOSFET, as shown in Fig. 3, the impurity concentration Nb of the region P2 contributes to Vth, and the profile thereof is such that Vtho decreases as the top silicon layer becomes thicker. On the other hand, the impurity concentration Na of the regions P1 and P3 also contributes to Vth, and the profile thereof is such that Vth_edge increases as the top silicon layer becomes thicker. Thus, the total Vth of the SOI MOSFET is almost constant.
As described above, the SOI MOSFET of the present invention can suppress the dependency on the thickness of the top silicon layer. Also since the impurity concentrations are set as Na > Nb, the short-channel effect and the punch-through can be reduced.
The SOI MOSFET having these characteristics may be produced as follows.
As a SOI substrate 4, is used a silicon substrate 1 on which a buried oxide film 2 of about 120 nm thickness and a top silicon layer 3 of about 50 nm thickness are formed in this order. The active region of the MOSFT is defined by forming a device isolation film 5 by a LOCOS method. The top silicon layer 3 has a thickness which allows operation in fully depleted mode.
Next, as shown in Fig. 4(a), channel ion implantation (background implantation) is carried out on the active region of the MOSFET on the top silicon layer 3 (50 nm). In the case of a PMOSFET, the ion implantation is performed at an implantation energy of 10 keV at a dose of phosphorus ions 6 of about 1 to 4 × 10¹² ions/cm², for example. Thereby, a channel central region P2 can be formed which has an impurity concentration profile (at implantation) as shown in Fig. 5, and it is possible to obtain a substantially uniform final impurity concentration (about ~5×10¹⁶ ions/cm³) in the horizontal direction in the channel central region. The implantation energy satisfies Rp/T_si ≒ 1/4 since the projected range Rp is about 14 nm.
Next, as shown in Fig. 4(b), a gate insulating film 7 is formed on the entire surface of the tope silicon layer 3, and a gate electrode 8 with a channel length of about 0.18 µm is formed by a usual process. Thereafter, using the gate electrode 8 as a mask, tilt ion implantation is carried out on the channel edges. The tilt ion implantation is performed at a tilt angle θ of about 30°, an implantation energy of about 70 to 90 keV, a dose of phosphorus ions 9 of about 1 to 3 × 10¹² ions/cm² in two steps (by rotation).
Thereby, an impurity concentration profile as shown in Figs. 2(a) and 2(b) is obtained at the channel edges. In the impurity concentration profile shown in Fig. 2(a), a dotted line represents the impurity concentration profile (about 5 to 6 × 10¹⁷ ions/cm³) at implantation, and a solid line represents the final implantation concentration profile (about 3×10¹⁷ ions/cm³). The channel edge regions are formed to have a length La = 0.06 µm approximately. The length La is determined to satisfy conditions for full depletion in consideration of process margins.
Thereafter, as shown in Fig. 4(c), BF₂ ions are implanted using the gate electrode 8 as a mask at an implantation energy of about 20 keV at a dose of about 4 × 10¹⁵ ions/cm² to form source/drain regions 10.
Thus the SOI MOSFET shown in Fig. 1 is produced.
In the above-described example, the tilt ion implantation is similar to that disclosed in USP 5841170 for forming a non-uniform impurity channel. Accordingly, the final impurity concentration profile by the tilt ion implantation can reduce the short channel effect and the punch through as described in USP 5841170 . Further, by optimizing both the cannel ion implantation and the ion implantation into the channel edges, it is possible to reduce fluctuations in the electrical characteristics of transistors due to variations in the thickness of the top silicon layer.
With regard to the channel ion implantation, Fig. 6 shows relationships between the thickness T_si of the top silicon layer and the total threshold voltage Vth in the cases where the ion implantation is performed by varying the thickness of the top silicon layer and the ion implantation energy. In Fig. 6, phosphorus ions are implanted at implantation energies of 10 keV (represented by a solid line), 20 keV (represented by a dotted line) and 40 keV (represented by an alternate long and short dash line). Fig. 7 shows a relationship between the change of the threshold voltage and the change of the thickness of the top silicon layer, i.e., Δ Vth/ Δ T_si of the SOI MOSFET, as a function of the thickness of the top silicon layer. In Fig. 7, ions are implanted at implantation energies of 40 keV, 30 keV, 20 keV and 12 keV.
According to Fig. 6 and Fig. 7, if the implantation energy is 40 keV, since Rp is about 49 nm, the change of the threshold voltage to the thickness of the top silicon layer (Δ Vth/ Δ Tsi) is the largest, about 17 mV/nm. On the other hand, if the implantation energy is small (i.e., Rp is small), Δ Vth/ Δ T_si becomes almost 0. Especially, if Rp is one-fourth or less of the thickness of the top silicon layer, Δ Vth/ Δ T_si is negative.
Further, by varying the impurity ions implantation conditions for and the thickness of the top silicon layer, the change of the total threshold voltage Vth of the SOI MOSFET is observed with respect to the thickness of the top semiconductor layer. The results are shown in Figs. 8(a) and 8(b).
According to Fig. 8(a), in the case where the SOI MOSFET is formed by a usual production method, Δ Vth/ Δ T_si is about 18 mV/nm, which is a large value, at E = 40 keV (solid circles in Fig. 8(a)).
On the other hand, if Rp/T_si is about 0.6, Δ Vth / Δ T_si is improved to about 7 mV/nm at E = 25 keV (hollow circles in Fig. 8(a)).
As shown in Fig. 8(b), if Rp ≒ TSi/4, Δ Vth/ Δ T_si is about 0.2 mV/nm at E = 12 keV. The Vth fluctuations with respect to variations in the thickness of the top semiconductor layer can be suppressed.
According to the present invention, the threshold voltage Vtho of the channel central region and the threshold voltage Vth_edge of the channel edge regions are so set that the change of Vtho with respect to the change of the thickness of the top semiconductor layer and the change of the Vth_edge to the change of the thickness of the top semiconductor layer are of opposite sign. Accordingly, the short-channel effect and the punch through can be effectively reduced, while the influence of the thickness of the top semiconductor layer on the electrical characteristics can be reduced, which leads to the production of highly reliable SOI MOSFETs.
Especially, the change of the threshold voltage owing to variations in the thickness of the top semiconductor layer can be effectively suppressed in the case where the changes of the threshold voltages of the channel central region and the channel edge regions with respect to the change of the thickness T_si of the top semiconductor layer meet sign(Δ Vth₀/ Δ T_si) < 0 and sign(Δ Vth_edge/ Δ T_si) > 0, or sign(Δ Vth₀/ Δ T_si) > 0 and sign(Δ Vth_edge/ Δ T_si) < 0 and the change of the threshold voltage Vth of the entire channel region with respect to the change of the thickness T_si of the top semiconductor layer meets Δ Vth/ Δ T_si) ≒ 0; or in the case where the channel central region is formed to have a peak concentration at the projected range Rp, which has a depth of half or less of the thickness T_si of the top semiconductor layer, by implanting ions of a first conductivity type and the channel edge regions have a constant impurity concentration in the depth direction of the top semiconductor layer and has a lateral length of one-third or less of the minimum channel length; or in the case where the channel central region has a constant impurity concentration in the depth direction of the top semiconductor layer, the channel edge regions are formed to have a peak concentration at a projected range Rp which has a depth of half or less of the thickness T_si of the top semiconductor layer, by implanting ions of a first conductivity type and the channel edge regions has a lateral length of one-third or less of the minimum channel length.
Since the method of the present invention is so compatible with usual production methods for semiconductors that fluctuations of the threshold voltage can be suppressed and production margins and yield can be improved without adding complicated production steps. Further, the decrease of the fluctuations of the electrical characteristics increases the operational margins of devices and simplifies the design of the devices. Consequently, the production process can be simplified and the production costs are reduced.

Claims

A method for producing a SOI MOSFET which includes a fully depleted channel region of a first conductivity type formed in a top semiconductor layer disposed on an insulative substrate, source/drain regions of a second conductivity type formed to sandwich the channel region and a gate electrode formed on the channel region with intervention of a gate insulating film, the method comprising:
providing, by implanting ions of a first conductivity type, a channel region comprising a channel central region and channel edge regions so that the impurity concentration of the channel edge regions adjacent to the source/drain regions is higher than the impurity concentration of the channel central region and further that changes of the threshold voltages of the channel central region and the channel edge regions with respect to the change of the thickness T_Si of the top semiconductor layer meet $sign (Δ {Vth}_{0} / Δ T_{Si}) < 0 and sign (Δ {Vth}_{edge} / Δ T_{Si}) > 0.$

and a change of a threshold voltage Vth of the channel region as a whole meets $(ΔVth / Δ T_{Si}) \approx 0 \frac{mV}{nm}$

by
forming the channel central region to have a peak concentration below the interface between the channel and the gate insulating surface at a projected range Rp, which has a depth of half or less of a thickness T_Si of the top silicon layer, such that a positive gradient of the concentration is achieved at the semiconductor surface when viewed in a depth direction from the interface side; and

forming the channel edge regions to have an almost uniform impurity concentration in the depth direction and the horizontal direction and to have a lateral length of one-third or less of the minimum channel length.
A method for producing a SOI MOSFET which includes a fully depleted channel region of a first conductivity type formed in a top semiconductor layer disposed on an insulative substrate, source/drain regions of a second conductivity type formed to sandwich the channel region and a gate electrode formed on the channel region with intervention of a gate insulating film, the method comprising:
providing, by implanting ions of a first conductivity type, a channel region comprising a channel central region and channel edge regions so that the impurity concentration of the channel edge regions adjacent to the source/drain regions is higher than the impurity concentration of the channel central region and further that changes of the threshold voltages of the channel central region and the channel edge regions with respect to the change of the thickness T_si of the top semiconductor layer meet $sign (Δ {Vth}_{0} / Δ T_{Si}) > 0 and sign (Δ {Vth}_{edge} / Δ T_{Si}) < 0$

and a change of a threshold voltage Vth of the channel region as a whole meets $(ΔVth / Δ T_{Si}) \approx 0 \frac{mV}{nm}$

by

forming the channel central region to have a constant impurity concentration in the depth direction of a top silicon layer; and

forming the channel edge regions to have a peak concentration below the interface between the channel and the gate insulating surface at a projected range Rp which has a depth of half or less of a thickness T_si of a top silicon layer, such that a positive gradient of the concentration is achieved at the semiconductor surface when viewed in a depth direction from the interface side, and to have a lateral length of one-third or less of the length of the channel region.
The method of claims 1or 2, wherein the formation of the channel region comprises forming, by implanting ions of the first conductivity type, so that the impurity concentration Nb in the channel central region and the impurity concentration Na in the channel edge regions meets Na/Nb = 3 to 6.